Caleb,
Your multi-camera setup does not look too rigid. Small changes in the
positioning of a camera would cause havoc on the final result.
With a multi-camera rig there will always be some parallax error. When shooting
video with the rig moving around the sometimes the subject will be close
On Sat 21-Jul-2012 at 10:29 -0300, Jim Watters wrote:
With a multi-camera rig there will always be some parallax error.
When shooting video with the rig moving around the sometimes the
subject will be close sometimes at infinity.
I think Caleb's intention is to use this rig in a place where
I got this result last year:
http://www.youtube.com/watch?v=u6DMSjtrTP0
You need to get the seams right first.
Be aware that a template will work for a single distance only.
So you need to know what kind of situation you will shoot later, and make a
template for that exact situation.
E.g. ALL cont
So many good pieces of info from everybody. Thanks.
I'm uploading a new video that uses some of the input you've provided. The
no-optimize helped a lot with the walking blending issue. I even generate
the blend masks on the first input file and use it on all subsequent files.
Presumably that sa
My camera rotates. I'd like to try to remove the rotation. I've got a
couple systems setup to help me retrieve the rotation data from frame to
frame. I am wondering how to apply that data pre-stitching, rather than my
current method of post-stitching. I noticed when I move the center of the
sti
On Sun 29-Jul-2012 at 10:50 -0700, Caleb wrote:
My camera rotates. I'd like to try to remove the rotation. I've got a
couple systems setup to help me retrieve the rotation data from frame to
frame. I am wondering how to apply that data pre-stitching, rather than my
current method of post-stitchin
On Sunday, July 29, 2012 3:55:26 PM UTC-5, Bruno Postle wrote:
>
>
> Yaw is simple, you just increment the y values for each photo in the
> project file. For any combination of roll, pitch and yaw, there is
> a transform-pano tool in the Panotools::Script perl module that can
> do the calcula
I was hoping I'd have better results than I am. I think I require more
understanding.
I have several 360x180 images stitched from six cameras. The cameras were
stable relatively, but rotated globally.
I want to correct the rotation in all images so as to match the first in
camera orientation.
I
Did I get caught in a spam filter? I replied to this thread last night and
it still hasn't shown up.
--
You received this message because you are subscribed to the Google Groups
"Hugin and other free panoramic software" group.
A list of frequently asked questions is available at:
http://wiki.p
Hi Caleb,
I don't know if I understood half of what you said, as English is also not
my native language, but I'll try to say some things to help.
1. Pay attention to your control points. With those images I would set up
only manual control points, as the automation can get confused with so many
e
On Sun 25-Nov-2012 at 17:43 -0600, Caleb Anderson wrote:
I was hoping I'd have better results than I am. I think I require more
understanding.
I have several 360x180 images stitched from six cameras. The cameras were
stable relatively, but rotated globally.
I want to correct the rotation in all
On Mon, Nov 26, 2012 at 3:41 PM, Bruno Postle wrote:
>
> ..but it is otherwise ok?
Well, no. I'm not sure what you mean by otherwise ok, but the whole goal
with that was to get the image aligned to the first image, so it's not
really ok. Unless you mean that it's still a properly stitched image,
I gave that a quick shot with some quickly placed control points. It gives
some really weird results when using 360x180 source images. It definitely
did not want me to keep the images in one stack.
http://res0l.net/StratoSphere/stacked.tar.gz contains my pto, inputs, and
sample outputs. Did I do s
On Mon, Nov 26, 2012 at 4:59 PM, Caleb Anderson
wrote:
> I gave that a quick shot with some quickly placed control points. It gives
> some really weird results when using 360x180 source images. It definitely
> did not want me to keep the images in one stack.
>
> http://res0l.net/StratoSphere/sta
On Wed 28-Nov-2012 at 11:25 -0600, Caleb Anderson wrote:
I do get better results if I only link the first image to all subsequent
images, instead of each image with each other. That actually makes sense
when I think about it.
However, I still get the strange lens/projection issue, even when jus
I added my images and, in the popup on add, set to equirectangular and hfov
of 360. Then I added control points and used the optimize tab to optimize
positions only.
Nona with gpu refused to output remapped images. Turning off gpu output
'properly'. (that is, the gl preview did show the same poorl
On Wed, Nov 28, 2012 at 3:12 PM, Caleb Anderson wrote:
> Nona with gpu refused to output remapped images. Turning off gpu output
> 'properly'. (that is, the gl preview did show the same poorly projected
> output)
Err, forgot to finish my thought. Turning off gpu output spit out
"properly" remapp
On Wed 28-Nov-2012 at 19:09 -0600, Caleb Anderson wrote:
Err, forgot to finish my thought. Turning off gpu output spit out
"properly" remapped images. (the broken images you see in the
video)
If I understand correctly, I don't have much as far as lens
distortion parameters in regards to a 3
18 matches
Mail list logo
